Getting the Social Cost of Carbon Right

Quantifying the global economic impacts of climate change is an enormously complex and high-stakes task that demands due diligence.

By Chris Warren

On January 20, 2021, newly inaugurated President Joe Biden signed 17 executive orders covering topics ranging from immigration to the COVID-19 pandemic to climate change. Among the orders issued on that first day of the new administration was one focused on the social cost of carbon.

The social cost of carbon is a monetary estimate attempting to quantify the global economic impact of emitting just one ton of the greenhouse gas carbon dioxide. Social cost of carbon estimates first gained relevance in the federal government during the George W. Bush administration, when the U.S. Ninth Circuit Court of Appeals ruled that the National Highway Transportation Safety Administration had been arbitrary and capricious in not valuing the benefits of greenhouse gas emissions reductions in setting vehicle efficiency standards.

The ruling kicked off an effort within the federal government to develop a methodology for determining the social cost of carbon. “The ruling that came back from the Ninth Circuit meant that the Department of Transportation had to come up with a number to use in their vehicle efficiency standards setting analysis,” said Steve Rose, a principal research economist at EPRI for 15 years, whose work focuses on long-term modeling of socioeconomic systems, climate change, and climate change impacts. “It also meant that every rule affecting greenhouse gas emissions from every agency would need to value expected emissions changes.”

At the time, Rose worked at the U.S. Environmental Protection Agency (EPA) and was researching methodologies to calculate the social cost of carbon. As one of the only people in the federal government studying the topic, Rose was called upon to educate government officials and technical staff on these methodologies and potential estimates.

The original rationale for developing estimates was to inform federal rulemaking. “The estimates were specifically developed for regulatory use and policies that only change global emissions incrementally. Agencies propose regulations, such as the EPA’s Clean Power Plan under the Obama administration, energy efficiency standards out of the U.S. Department of Energy, or even U.S. Department of Agriculture policies with respect to agricultural practices,” Rose said. “In all of the cases, the idea was that if the proposed regulation affected greenhouse gas emissions, then you needed to value them and include that as part of your cost-benefit analysis.”

Current Methodology Has Been in Place for a Decade-Plus

It wasn’t until 2010, during the Obama administration, that an interagency working group from across the federal government established a methodology to calculate standardized estimates of the social cost of carbon that all federal agencies could use. This 2010 methodology is still in use today, and the Biden administration decided to use it to generate its “interim” estimates. The interim estimates have remained in force while the administration works on developing updated and improved estimates, as directed by the president’s executive order. The order requested interim estimates, development of updated estimates, and reconstituted the working group, which had been disbanded during the Trump administration.

On November 11, 2022, as part of a proposed rule to reduce methane emissions in the oil and gas industry, the EPA released a draft new methodology and proposed updated social cost of carbon estimates, as well as social cost estimates for other greenhouse gases—methane and nitrous oxide. While the EPA’s draft new estimates have not been officially declared the interagency working group’s proposed revised estimates, they are being considered.

If anything, the need for scientifically rigorous development of the social costs of carbon and other greenhouse gases has increased since 2010. That’s because their use has expanded well beyond federal regulations and has the potential to influence a growing number of investments and decisions. For example, there are proposals to incorporate the social cost of carbon into the environmental impact analyses required under the National Environmental Protection Act. Other proposals seek to incorporate the metric into federal government procurement decisions by valuing the greenhouse gas emissions associated with potential suppliers’ products and services.

The social cost of carbon is also being used, or is proposed for use, at the state level in regulatory analyses and other contexts, such as externalities pricing in resource planning and power dispatch. For instance, when a utility submits a generation resource plan to its state public utilities commission (PUC), the PUC may ask the utility to value the greenhouse gas emissions associated with the proposed investments.

Dispatch pricing is another potential power sector application for the social cost of carbon. “There are a variety of issues here, but as system operators are dispatching different resources, they could consider including the cost of the greenhouse gas emissions of each resource. So, if the system operator is choosing between a renewable resource with no emissions and a gas plant with emissions, they might include an adder on the gas plant dispatch price,” Rose said. “However, it is more complicated than that, especially if the plant’s emissions are already being regulated, or there are different emissions policies across the states in the system.”

Recommendations to Improve the Methodology

Estimating the social cost of carbon is important, but it’s also an enormous scientific challenge. For example, carbon dioxide released by burning fossil fuels, as well as other activities, remains in the atmosphere for a hundred years. It also impacts how the world stores and releases carbon beyond that. “Which means that estimating the social cost of carbon requires modeling the world’s physical and economic systems centuries into the future,” says Rose. “Given the need to quantify future global economies and climate change for centuries, and how climate change could affect everything from agriculture to human health to power systems to coastal infrastructure, there is significant uncertainty that needs to be considered, quantified, and incorporated into the modeling.”

The complexity of developing a social cost of carbon methodology and estimates underscores the importance of approaching the task with scientific rigor and transparency. Soon after the administration released its executive order in January 2021, EPRI published a technical report, Repairing the Social Cost of Carbon Framework, that laid out a to-do list for updating the “interim” methodology in use since 2010. It included immediate fixes to the interim approach and specific scientific issues that needed to be overcome by any new methodology.

EPRI’s analysis drew on over a dozen years of study of the social cost of carbon methodologies, as well as EPRI’s participation on the National Academies of Sciences, Engineering, and Medicine (NASEM) social cost of carbon committee, which the Obama administration had assembled to recommend methodological improvements. In its executive order, the Biden administration called out the NASEM recommendations explicitly, stating that they should be considered in developing a new methodology and estimates.

The EPA’s recent draft new methodology, however, does not address many of the NASEM recommendations. At a foundational level, the draft methodology is not scientifically reliable and robust. “What that means is that it lacks transparency and justification, and it isn’t taking into account the broader scientific knowledge available, including not properly capturing uncertainty. Therefore, the numbers are not robust and would change if we used equally relevant and plausible alternative specifications and assumptions,” Rose said. “The draft new estimates are simply not numbers we can rely on to influence billions of dollars of decisions.”

For example, NASEM, in its Phase 1 study, noted that the interim estimates and the methodology they were based on needed numerous revisions and that a partial fix would be insufficient; there were simply too many problems that needed to be addressed. Therefore, in its Phase 2 study, NASEM provided comprehensive recommendations for revisions to the framework overall and individual modules, which included fully considering the available science and uncertainties.

EPRI’s recent analysis of the EPA’s draft new methodology and estimates found that the NASEM recommendations were not addressed. EPRI’s analysis also found that the methodology has substantive scientific issues with, for instance, robustness, plausibility, discounting, and transparency, in addition to a lack of internal consistency and coherency across the modeling.

Taking the Time to Ensure Scientific Due Diligence

EPRI’s assessment of the proposed new methodology is meant to help the public understand the approach and relevant science. “Our primary function is facilitating understanding and productive conversation that helps in the eventual development of a scientifically reliable methodology and set of estimates,” Rose said.

There are opportunities to enhance scientific rigor, transparency, and public trust regarding the EPA’s draft new social cost of carbon methodology and estimates. Steps that could be taken include assembly of a peer review panel that appropriately evaluates the methodology. While a peer review panel was assembled, it was not asked to assess the scientific reliability of the methodology or provide feedback and recommendations agreed to by all the panelists. It was also limited in size and expertise and in its ability to engage in debate and dialogue and consider public input. Such dialogue and input are good scientific practices and valuable for engendering public confidence.

Because transparency is important to understanding and assessment, the EPA’s draft methodology documentation should be enhanced to clearly communicate and justify the details. These would include the data, equations, parameters, uncertainty specifications, sources used, and detailed results that establish the methodology’s reliability and robustness at every step in the calculations.

Producing a scientifically robust social cost of carbon is complex and challenging. Research and dialogue, however, also need to extend to how social costs of greenhouse gas estimates are eventually applied. “A whole new set of technical issues arise when we start talking about applications,” Rose said. “Thus, this is not just a conversation about how you get the best estimates for the social cost of carbon. This is also a conversation about how to apply them properly.”

EPRI has been evaluating and tracking applications of the social cost of carbon for over a decade. That line of research has found that the same carbon is being valued more than once across policies, from mineral extraction to fuel combustion to power dispatch to the use of goods and services. This multiple pricing of carbon and other greenhouse gases increases the cost of reducing carbon emissions without further reducing those emissions. EPRI’s work has also found important inconsistencies across calculations used in benefit-cost analyses, as well as carbon emissions increases, commonly called leakage, elsewhere in the economy. Together, these affect the climate and net benefit estimates of policies and their reliability. For example, EPRI has found these issues to be present in the benefit-cost analyses for EPA’s recent proposed oil and gas methane and 111 power plant rules. “Overall, these technical issues, along with the social cost of carbon estimation issues, undermine confidence in the application results and insights that are informing decisions,” said Rose.

While the current administration has a sense of urgency to revise the social cost of carbon methodology and estimates, the process to date has not provided the necessary scientific due diligence. “The interim estimates methodology has been with us for 13 years. We need to get things right. Whatever methodology we are going to have next is probably going to be with us for a long time,” Rose said. “We need to follow proper scientific process in order to create estimates and insights that are scientifically reliable, robust, and stable. There is still much to do to get there, but fortunately, there is a clear sound scientific path forward and specific opportunities for improving both the process and the scientific basis.”

---